Created attachment 20921 [details]
The hyperspace demo from rss-glx (Really Slick Screensavers) causes an immediate hang in X.
Killing hyperspace does not help and the hardware is wedged upon killing Xorg.
-- chipset: G45 / ICH10R
-- system architecture: 32-bit
-- xf86-video-intel: bea98cdfd93fc1181a06c51e57fcab227ff4827e
-- xserver: 1.5.2
-- mesa: a0d5c3cfe6582f8294154f6877319193458158a2
-- drm: c99566fb810c9d8cae5e9cd39d1772b55e2f514c
-- kernel: for-airlied 66647dc60d16fae9f6963fd98b6d9baa1a8dac69
-- Linux distribution: Debian unstable
-- Machine or mobo model: Asus P5Q-EM
-- Display connector: DVI
Created attachment 20922 [details]
Backtrace from hyperspace
Created attachment 20923 [details]
Backtrace from Xorg
reproduced on g45. Noted that the app only sets s/t wrap modes on cube mapping, but the docs claim that we need to have all 3 in the same mode or it gets angry. Smashing all 3 settings to CLAMP didn't help, though. The hang is on what is, at first glance, a perfectly healthy batchbuffer. (INTEL_DEBUG=bat,sync)
I have a very similar issue on a box with a GA-EG45M-DS2H (G45) motherboard. The backtrace I see in xorg is different though and I also don't have any "[mi] EQ overflow" spam in my xorg.log. Once there is a patch for this, please post the commit sha1 or .patch and I will test it on my box as well.
My bug is here:
*** Bug 20086 has been marked as a duplicate of this bug. ***
I can no longer reproduce this with Mesa master. Can anyone else test?
If would be useful to understand when this got fixed (for potential cherry picking), I might try to bisect later (if I find the time) but if anyone got known good or known bad commits, please paste them here.
I just tried Mesa 2305642b2e8edcebdc727f1181f7dbfcc78e8028 and it no longer causes a hang.
Mass version move, cvs -> git